The industry is spending its legal budget chasing a startup. Meanwhile, billions in royalties sit undelivered in the infrastructure that was supposed to protect creators. Nobody wants to talk about that.
The industry is having the wrong argument.
Suno just raised $250 million and hit a $2.45 billion valuation. Everybody has something to say about it. The lawsuits, the training data, the copyright implications, and the democratisation of music creation. It is a legitimate conversation.
It is just not the most urgent one.
The most urgent one is: why does the global royalty black box sit at an estimated $2.5 billion, and why is nobody in a conference room talking about it with the same energy they spend on Suno?
That is not a rhetorical question. The answer tells you a lot about who benefits from keeping things the way they are.
Suno is a visible enemy with money and lawyers. You can sue Suno. You can put Suno in a headline. The structural failures of royalty collection implicate too many of the people who would need to fix them.
The numbers
The Mechanical Licensing Collective in the US received $424 million in historical unmatched royalties from digital service providers. Apple Music alone transferred over $163 million of that. Spotify has over $152 million. These were not disputed royalties. They were royalties that could not be delivered because nobody could figure out who owned the music.
That is just the documented slice of a much larger problem, and it only covers one country for one historical period.
Global estimates put the black box at $2.5 billion across PROs, CMOs and DSPs worldwide. And it compounds every year, because the catalogue grows faster than the infrastructure designed to track it.
The causes are not mysterious. A recording released without a linked ISRC. A composition without an ISWC, making it invisible to international PRO matching. A publisher not registered in a territory where the song streams 40,000 times a week. Metadata that breaks somewhere between the distributor and DSP and the collection society. Split sheets that live in somebody’s email drafts and nowhere else.
Each gap is individually small. Together, they are a systemic failure that the industry has managed for decades by not measuring it too carefully.
Why does this stay broken
The black box is not an accident.
When unmatched royalties cannot be returned to owners within the holding period, most collection societies distribute them to registered members on a market-share basis. This means the money flows to whoever already earns the most. A PRO where five major publishers dominate the membership will redistribute unmatched money from independent and international rights holders into those publishers’ pockets.
This is not corruption in the traditional sense. It is a distribution rule. But it creates a structural disincentive to improve matching rates. Every percentage point of improvement means less money flowing through market-share distribution, which means less going to the largest stakeholders. Those stakeholders also tend to sit on the boards of the organisations responsible for improving the matching.
Nobody is going to publicly press for accountability on that. Not the organisations. Not the labels that benefit from the current flow. Not the DSPs, who have no contractual obligation to chase the reconciliation on your behalf.
Suno, by contrast, is a clean target. New money. No institutional relationships to protect. An actual legal theory to argue. The camera stays on Suno because pointing it elsewhere is expensive.
The companies are actually working on it
A handful of companies are building into the problem rather than around it. They operate at a fraction of the capital and attention that generative AI receives.
Music Reports acquired Blokur, a company whose matching technology is built on graph theory to identify unclaimed publishing royalties at scale. The resulting product, Trakdex, is part of the MLC Supplemental Matching Network. Music Reports operates without membership bias, which means its incentives are aligned with accurate matching rather than with any particular stakeholder.
BMG partnered with Google Cloud to build StreamSight, an AI tool that uses machine learning to analyse royalty data and flag anomalies, including missing sales periods and mismatches between reported revenue and rights ownership. A major label is investing in the transparency layer. That is worth noting.
Vollou is solving the live performance side, which is arguably more broken than digital. PRS for Music data cited by the company shows that only 10 out of 230 DJs submitted their setlists at a major British festival. Vollou captures setlists automatically through music recognition, generating the verified data that PROs need to distribute performance royalties correctly.
Notes.fm automates registration at the point of release. Artists in their beta found that roughly a quarter of their catalogue had missing or incomplete registrations. Some recovered six figures in unclaimed royalties from works already in distribution.
These are not glamorous products. They solve problems that are embarrassing to acknowledge publicly, because acknowledging them means admitting the system failed.
These companies are finding money that already existed and was not delivered. That is a far more uncomfortable finding for the industry than anything Suno is doing.
The data that already exists, and most people are not using it
There is a separate conversation here, and it is a more optimistic one.
The black box is a structural problem protected by institutional incentives. But there is another category of uncollected royalties that has nothing to do with broken infrastructure. It has to do with not reading the data that is already available.
Platforms like Spotify for Publishers give rights holders access to detailed usage data across their catalogue. Stream counts by territory, by track, by time period. When you cross-reference that data against what your collection society actually paid, patterns emerge. Works generating consistent volume in territories where you have no registered representation. Tracks are performing well on a platform that does not appear in your statement at all. Discrepancies between what was played and what was reported point directly to registration gaps you can fix.
This is not a black box situation. The data is there. The tools exist. The gap is analytical, not structural.
Rights holders who understand how to read these patterns and act on them recover royalties that others leave on the table permanently. Not because the system failed them. Because they never looked.
The difference between a rights holder who collects what they are owed and one who does not is increasingly a data literacy problem, not a rights problem.
What Water and Music has been tracking
Water and Music, the research platform founded by Cherie Hu, has spent over a decade being one of the few voices willing to map the unsexy infrastructure of the music business with real rigour. Their State of Data report surveys how professionals actually use data. Their investment database tracks over 700 companies.
Their read on 2024 music tech investment was direct: the biggest checks went to the industry’s least glamorous problems. Infrastructure over spectacle. Operational reality over hype.
They also flagged something important about Suno’s Series B. Pulling the Crunchbase data showed it was the largest music tech investment in years. Not just a company milestone. A macro signal about where capital thinks value creation in music is happening. And it is not in the royalty infrastructure.
What accountability would actually look like
If the industry were serious about creator compensation, it would treat data infrastructure with the urgency it applies to copyright litigation.
That means publicly reporting matching rates by territory. Requiring DSPs to provide reconciliation reports that connect analytics to payments. Funding metadata standards with the same aggression used to pursue AI training data licenses. Disclosing exactly how black box distributions are calculated and who receives them.
None of this requires new legislation. None of it requires waiting for Suno to settle. Most of it requires only that the institutions responsible for collecting and distributing creator royalties publish the numbers they already have.
The fact that they do not is itself an answer.
The music industry is not afraid of data. It is afraid of specific data. The kind that shows the gap between what was collected and what was paid. Between what was played and what was matched. Between what the analytics say happened and what the statement says was earned.
Suno is a much more comfortable place to look.